mbox series

[0/3] src/feature: add IO_URING feature checking

Message ID 20200916122327.398-1-zlang@redhat.com (mailing list archive)
Headers show
Series src/feature: add IO_URING feature checking | expand

Message

Zorro Lang Sept. 16, 2020, 12:23 p.m. UTC
This patchset bases on https://patchwork.kernel.org/cover/11769847/, which
makes xfstests fsstress and fsx supports IO_URING.

The io_uring IOs in fsstress will be run automatically when fsstress get
running. But fsx need a special option '-U' to run IO_URING read/write, so
add two new cases to xfstests to do fsx buffered and direct IO IO_URING
test.

[1/3] new helper to require io_uring feature
[2/3] fsx buffered IO io_uring test
[3/3] fsx direct IO io_uring test

And the [2/3] just found an io_uring regression bug (need LVM TEST_DEV):
https://bugzilla.kernel.org/show_bug.cgi?id=209243

Feel free to tell me, if you have more suggestions to test io_uring on
filesystem.

Thanks,
Zorro

Comments

Zorro Lang Sept. 17, 2020, 4:05 a.m. UTC | #1
On Wed, Sep 16, 2020 at 08:23:24PM +0800, Zorro Lang wrote:
> This patchset bases on https://patchwork.kernel.org/cover/11769847/, which
> makes xfstests fsstress and fsx supports IO_URING.
> 
> The io_uring IOs in fsstress will be run automatically when fsstress get
> running. But fsx need a special option '-U' to run IO_URING read/write, so
> add two new cases to xfstests to do fsx buffered and direct IO IO_URING
> test.
> 
> [1/3] new helper to require io_uring feature
> [2/3] fsx buffered IO io_uring test
> [3/3] fsx direct IO io_uring test
> 
> And the [2/3] just found an io_uring regression bug (need LVM TEST_DEV):
> https://bugzilla.kernel.org/show_bug.cgi?id=209243
> 
> Feel free to tell me, if you have more suggestions to test io_uring on
> filesystem.
> 
> Thanks,
> Zorro

Err... Sorry about 3 duplicate patchset. I tried to sent this patchset 3 times
yesterday, but all failed, and I got feedback from server:
  "The message you sent to fstests@vger.kernel.org hasn't been delivered yet due to: Communications error."

I thought these emails have been abandoned, never know they're delayed. Please
only review the last patchset of these duplicate things.

Thanks,
Zorro

>